ResearchTasksNumerical Question AnsweringNumerical Question AnsweringFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedFinQA (test)Human Expert Performance91.16Execution Accuracy33Feb 27, 2026FinLongDocQA 1.0 (test)FinLongDocAgent41.34EM19Apr 7, 2026FinQA 1.0 (test)Human Expert Performance91.16Execution Accuracy14Feb 26, 2026MathQA (test)ELASTIC83Program Accuracy6Feb 27, 2026